A systematic scoping review of the methodological characteristics relating to subgroup analyses within universal school-based social and emotional learning interventions: a critical appraisal of the evidence and recommendations for the field

TY - JOUR

T1 - A systematic scoping review of the methodological characteristics relating to subgroup analyses within universal school-based social and emotional learning interventions

T2 - a critical appraisal of the evidence and recommendations for the field

AU - Hamilton, Suzanne

AU - Boehnke, Jan R.

AU - Qualter, Pamela

AU - O’Brien, Annie

AU - Humphrey, Neil

PY - 2026/9/1

Y1 - 2026/9/1

N2 - Background Universal school-based social and emotional learning (USB SEL) interventions may not benefit all students equally. Credible subgroup analyses are needed to understand differential effects and equity implications. This scoping review examined how subgroups are defined, analysed, reported, and prespecified in primary/elementary USB SEL trials. Methods Seven databases and trial registries were searched for reports dated from 2000 to August 2024. Eligible reports evaluated a USB SEL intervention using comparison/control-group designs reporting at least one subgroup analysis. Data were extracted on subgroup definitions and rationales, analytic and reporting practices, and trial registrations and protocols. Findings were narratively synthesised. Results Of 2,579 records, 118 reports representing 99 trials of 66 interventions were included. Gender and baseline difficulties were the most commonly examined moderators; other sociodemographic characteristics were examined less often, and analyses that considered multiple characteristics jointly were rare. Main-effect models accounted for clustering in 67.8% of reports; interaction-based tests were the most common subgroup approach (also 67.8%), but only two reports described adequate subgroup-analysis power and 16.1% adjusted for multiple comparisons. Subgroup effect sizes were reported in 72.9% and confidence intervals in 24.6%. Most analyses were described as planned and justified but 22.0% had a registration, 10.2% a protocol, and 11.9% prespecified all subgroup analyses in a located registration or protocol. Conclusions Although subgroup-analysis practice in USB SEL is improving, limited power, prespecification, demographic reporting, and transparency constrain inferences about differential effects. More consistent, theory-driven, prospectively documented, and adequately powered analyses are needed to determine whether benefits are equitably distributed. Impact statement • This scoping review synthesises how subgroup analyses are defined, conducted, reported, and prospectively documented in universal primary/elementary school-based SEL trials. • Across 118 reports, subgroup analyses most often examined gender and baseline levels of difficulty or skill, whereas differential effects for students from socioeconomically disadvantaged, ethnic minority, or different language backgrounds, students with special educational needs, and students with intersecting characteristics were examined comparatively rarely. •Although use of interaction-based and cluster-adjusted models has increased, subgroup analyses are rarely reported as adequately powered, adjustment for multiple comparisons is uncommon, and prospective documentation in registrations or protocols is limited. •The review offers practical recommendations to improve the transparency, statistical rigour, and equity relevance of subgroup analyses, supporting more credible conclusions about who benefits from universal SEL interventions.

AB - Background Universal school-based social and emotional learning (USB SEL) interventions may not benefit all students equally. Credible subgroup analyses are needed to understand differential effects and equity implications. This scoping review examined how subgroups are defined, analysed, reported, and prespecified in primary/elementary USB SEL trials. Methods Seven databases and trial registries were searched for reports dated from 2000 to August 2024. Eligible reports evaluated a USB SEL intervention using comparison/control-group designs reporting at least one subgroup analysis. Data were extracted on subgroup definitions and rationales, analytic and reporting practices, and trial registrations and protocols. Findings were narratively synthesised. Results Of 2,579 records, 118 reports representing 99 trials of 66 interventions were included. Gender and baseline difficulties were the most commonly examined moderators; other sociodemographic characteristics were examined less often, and analyses that considered multiple characteristics jointly were rare. Main-effect models accounted for clustering in 67.8% of reports; interaction-based tests were the most common subgroup approach (also 67.8%), but only two reports described adequate subgroup-analysis power and 16.1% adjusted for multiple comparisons. Subgroup effect sizes were reported in 72.9% and confidence intervals in 24.6%. Most analyses were described as planned and justified but 22.0% had a registration, 10.2% a protocol, and 11.9% prespecified all subgroup analyses in a located registration or protocol. Conclusions Although subgroup-analysis practice in USB SEL is improving, limited power, prespecification, demographic reporting, and transparency constrain inferences about differential effects. More consistent, theory-driven, prospectively documented, and adequately powered analyses are needed to determine whether benefits are equitably distributed. Impact statement • This scoping review synthesises how subgroup analyses are defined, conducted, reported, and prospectively documented in universal primary/elementary school-based SEL trials. • Across 118 reports, subgroup analyses most often examined gender and baseline levels of difficulty or skill, whereas differential effects for students from socioeconomically disadvantaged, ethnic minority, or different language backgrounds, students with special educational needs, and students with intersecting characteristics were examined comparatively rarely. •Although use of interaction-based and cluster-adjusted models has increased, subgroup analyses are rarely reported as adequately powered, adjustment for multiple comparisons is uncommon, and prospective documentation in registrations or protocols is limited. •The review offers practical recommendations to improve the transparency, statistical rigour, and equity relevance of subgroup analyses, supporting more credible conclusions about who benefits from universal SEL interventions.

U2 - 10.1016/j.sel.2026.100233

DO - 10.1016/j.sel.2026.100233

M3 - Article

SN - 2773-2339

SP - 100233

JO - Social and Emotional Learning: Research, Practice, and Policy

JF - Social and Emotional Learning: Research, Practice, and Policy

ER -